DSXchange

A reject link would work, or you can just believe the error message that row 53161308 has the issue. I'm guessing you've got an embedded high ascii character, probably a linefeed.

You're going to have to go thru the checklist and verify your installations. Do you have the requisite compiler installed, do you have the permissions setup, have you gone thru and created your node pools, etc.

Welcome aboard. Is this a new installation of PX?

It appears to be a dynamic array of directory path parameter names and their associated values. Your function appears to gather a list of tokens like this: Load_dir=/here/load Work_dir=/here/work Temp_dir=/here/temp etc. and copy the contents to an archive directory that the function creates.

Section 3-3, middle of the page of the DataStage Enterprise Edition Parallel Job Developer's Guide. Read it for yourself.

The difficulty in SCD in parallel is the complicated nature of a type 1, 2, and hybrid 1-2. You need to transform a new row and compare it to the current row. You may compare just a subset of all columns. This means that if you have 40 columns, maybe only 35 pertain to looking for a difference. You ...

Okay, I finally see what you are doing. The OCI_3-->Hashed portion of your job will run to completion before the first row is read from SRCFile. The Hashed file will be static while rows from SRCFile are passed to Link3 and Link2. The order of rows going into Link3---->OCI_3 will not be in sync or s...

Could you please edit your post and put code tags around the logic so I can read it?

Execution order is for an incoming row relative to the outgoing links. The incoming row with be put on the output links in the link execution order. Row buffering and IPC work at the row level, so they pertain to interaction with other stages. Your row will be written to the output link the correct ...

A Sequence with Command stages, a Batch job with DSExecute API calls, make your choice.

same way

Looks like you need to run the data to a hashed file with all rows flagged as "N" and also to an Aggregator getting the max for the group and then output from the aggregator to a transformer where you set the flag accordingly, then write to the same hashed file. Then, move the data out of the hashed...

A CONSTRAINT should be all that you need.

You have a job attempting to create a hashed file in the DS project that already exists.

ItsMe wrote:Well I don't have to do all these work if the AUTO-PURGE ever worked on multi-instance job! did not work for us... so we had to do it this way.

No choice@

You must be on 7.1 or earlier. This is now fixed under 7.5+.